Visual region editor for Krea 2 — boxes, per-box prompts and per-box LoRAs in a single node.
Press 'Edit layout' to open a detached editor window: drag and resize boxes on the canvas, give each one a prompt and an identity prompt, and attach any number of LoRAs per box with their own strength and routing mode. 'Backdrop' loads a recent render behind the canvas so boxes can be placed against a real composition.
Boxes are stored NORMALIZED (0..1) and rescaled shape-preserving when the canvas changes: a box keeps its width-to-height ratio and its relative centre instead of being squashed. That is the part plain fraction-based editors get wrong — there a square box silently turns into a wide rectangle when you go from 1:1 to 16:9.
Wire regions + loras + global_prompt + width + height straight into K2 Compose.
By ping1979ping·Created 5 months ago·Updated 21 days ago· 0
Output width. Changing it rescales the layout shape-preserving instead of distorting the boxes.
height
INT
102464–16384
Output height. Same shape-preserving rescale applies.
global_prompt
STRING
Scene ONLY — no people. Describe the subjects in their boxes instead.
Image tokens outside every region see only this text, so any person mentioned here gets drawn there. Measured with two regions: 'a sunlit park lawn with tall trees' → 2 faces, 'photo of a woman standing in a sunlit park' → 3, 'group portrait photo' → 5. Even the singular adds one too many.
